Intelligent wafer etching process

Abstract

The invention relates to an intelligent wafer etching process in the technical field of information, and the process comprises the steps: obtaining the etching rates of different materials under various liquid medicine formulas, constructing an etching rate database covering various material systems, screening out a material combination with high etching selectivity by comparing the etching rate differences in the database, and carrying out the etching of the material combination. Determining a targeted liquid medicine formula optimization direction; a multi-parameter online monitoring unit is installed in etching equipment, the concentration of liquid medicine is measured in real time through an optical fiber spectrometer, the temperature and the pH value of the liquid medicine are measured in real time through a thermocouple and a pH electrode, when monitoring parameters exceed a set range, a liquid medicine proportion adjusting program is started, and dynamic regulation and control over the property of the liquid medicine are achieved by controlling the liquid inlet amount and the liquid outlet amount. And the stability and the consistency of the etching process are ensured.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to an intelligent wafer etching process. Background Art

[0002] In the semiconductor etching process, improving etching uniformity and real-time control of chemical solution parameters are two closely related technical challenges. Traditional immersion etching is difficult to ensure uniformity of large-area wafers, and although spray etching can improve this problem, it still faces the problem of uneven etching in local areas.

[0003] The above problems are mainly caused by the uneven flow and diffusion of the solution on the wafer surface. At the same time, the concentration, temperature and pH of the solution will change over time during the etching process, affecting the etching rate and selectivity. How to design an efficient spray system and stirring device to ensure the uniform distribution of the solution on the wafer surface and monitor and control the solution parameters in real time is a key issue that needs to be solved urgently. This requires comprehensive consideration of factors such as the layout of the nozzle, the spray angle, the shape and rotation speed of the stirring blade, and at the same time establish a quantitative relationship between the solution parameters and the etching performance, and develop an intelligent online monitoring and feedback control system. Solving these interrelated technical problems is of great significance to improving the stability and consistency of the etching process. Summary of the invention

[0004] The present invention provides an intelligent wafer etching process, which comprises the following steps:

[0005] S101, obtaining etching rates of different materials under various chemical solution formulations, constructing an etching rate database covering various material systems, screening out material combinations with high etching selectivity by comparing etching rate differences in the database, and determining a targeted chemical solution formulation optimization direction;

[0006] S102, for the selected material combinations, use first-principles calculations to study the interaction mechanism between each component in the solution and the material surface, obtain the kinetic parameters of the surface reaction, perform molecular dynamics simulation, reveal the microscopic mechanism of the diffusion and mass transfer process of the solution, and establish a multi-scale coupled etching process model, including a molecular-scale surface reaction model, a mesoscopic-scale diffusion and mass transfer model, and a macroscopic-scale etching morphology evolution model. Through the coupling of these models, the rate-controlling steps and chemical reaction paths of material removal are clarified, guiding the optimization of the solution formulation;

[0007] S103, using microfluidic chip technology, designing and integrating multiple parallel etching reaction chambers to achieve high-throughput screening of drug solution formulas, forming a concentration gradient in the chip by controlling the flow rate of the sampling pump and the drug solution ratio, using spectroscopy to monitor the composition of the drug solution at the chip outlet in real time, obtaining etching rate data under different formulas, and using machine learning algorithms to establish a quantitative relationship between formula and etching performance, and preferably obtaining a drug solution formula with high etching selectivity;

[0008] S104, based on the screening results of the microfluidic chip, a spray head array and a stirring device are designed for the preferred liquid medicine formula to form a uniform liquid medicine spray and turbulent flow field. By optimizing the nozzle shape, the number of nozzles, the spray angle and the stirring blade parameters, the mass transfer process between the liquid medicine and the material surface is strengthened to improve the etching uniformity. The liquid medicine recovery device adopts temperature and conductivity dual parameter control to adjust the recovery time in real time according to the change of liquid medicine properties to reduce liquid medicine waste;

[0009] S105, installing a multi-parameter online monitoring unit in the etching equipment, using a fiber optic spectrometer to measure the concentration of the liquid in real time, using a thermocouple and a pH electrode to measure the temperature and pH of the liquid in real time, and when the monitoring parameters exceed the set range, starting the liquid ratio adjustment program, and realizing dynamic regulation of the liquid properties by controlling the liquid inlet and liquid discharge, to ensure the stability and consistency of the etching process;

[0010] S106, characterize and evaluate the etching results after dynamic control, using etching depth, surface roughness and etching residue as evaluation indicators, using optical microscopy and atomic force microscopy to characterize the etching results, and quantitatively evaluate the etching performance of different material systems and liquid formulations by calculating uniformity index and selectivity index. The uniformity index calculation formula is the ratio of the standard deviation of the etching depth to the average value, and the selectivity index calculation formula is the ratio of the etching rates of the two materials. For the performance evaluation results, gray correlation analysis is used to determine the key process parameters to guide the learning and iterative update of the intelligent optimization system;

[0011] S107, based on the process optimization experimental data, construct a nonlinear mapping model between the solution formula, process parameters and etching results. The model adopts a multi-layer neural network structure, with the input layer being the solution formula and process parameters, and the output layer being the etching depth, uniformity and selectivity indicators. The obtained etching laws and control strategies are transformed into dynamically updated knowledge reasoning rules, and an intelligent decision-making system based on case reasoning and fuzzy logic is developed to realize the automation and intelligence of solution formula design, process parameter selection and etching defect diagnosis.

[0012] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0013] The present invention discloses an intelligent wafer etching process, which constructs a multi-material etching rate database, combines first-principles calculations and molecular dynamics simulations, establishes a multi-scale coupled etching process model, uses microfluidic chip technology to achieve high-throughput screening of liquid medicine formulas, and uses machine learning algorithms to optimize formulas. The present invention designs dynamically controlled etching equipment, a spray head array, and a stirring device, and also includes a uniform spray system and a multi-parameter online monitoring unit, which realizes real-time control of the etching process. By characterizing and evaluating the etching results, a nonlinear mapping model is established, and an intelligent decision-making system based on case reasoning and fuzzy logic is developed. The present invention realizes the automation and intelligent optimization of the etching process, significantly improves the etching uniformity and selectivity, and provides an innovative solution for the etching process in microelectronics manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The present invention is a flow chart of an intelligent wafer etching process.

[0015] Figure 2 A schematic diagram of an intelligent wafer etching process of the present invention.

[0016] Figure 3 This is another schematic diagram of an intelligent wafer etching process of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] like Figure 1-3 In this embodiment, an intelligent wafer etching process may specifically include:

[0019] S101. Obtain the etching rates of different materials under various chemical solution formulations and build an etching rate database covering various material systems. By comparing the etching rate differences in the database, select material combinations with high etching selectivity and determine the targeted optimization direction of chemical solution formulations.

[0020] S101 also includes: obtaining etching rate data covering a preset number of types of material systems and liquid medicine formulas, and establishing the etching rate database; using a data analysis algorithm to analyze the etching rate data in the etching rate database, and calculating the etching rate differences of different material combinations under the same liquid medicine formula; judging the selectivity of the material combination according to the etching rate difference, and screening out the material combination with high etching selectivity; for the material combination with high etching selectivity, using a machine learning algorithm to analyze the relationship between the material combination and the liquid medicine formula parameters, and obtaining key liquid medicine formula parameters that affect the etching selectivity; determining the liquid medicine formula optimization direction according to the key liquid medicine formula parameters, and optimizing the liquid medicine formula by adjusting the value range of the key liquid medicine formula parameters; applying the optimized liquid medicine formula to the material combination with high etching selectivity, verifying the optimization effect through etching experiments, and if the preset optimization target is not achieved, returning to the liquid medicine formula optimization step, and continuing the optimization until the optimal liquid medicine formula is obtained to achieve high-selective etching of the material.

[0021] For example, we first need to clarify the different material systems and liquid formulations. Here are two examples of material systems. One is the silicon-based material commonly used in integrated circuit manufacturing, including single crystal silicon, polycrystalline silicon, silicon dioxide, silicon nitride, etc.; the other is a new two-dimensional material system, such as molybdenum disulfide, tungsten diselenide, etc. The liquid formulation depends on different materials and etching requirements. Common ones include acidic solutions, alkaline solutions, and solutions containing oxidants and chelating agents. For example, hydrofluoric acid solutions are often used to etch silicon dioxide, and potassium hydroxide solutions are used to etch single crystal silicon. These are all selective wet etching. Phosphoric acid, nitric acid and acetic acid are mixed in different proportions, and single crystal silicon, silicon dioxide and silicon nitride commonly used in the field of integrated circuits are placed in mixed solutions of different proportions, and etched at the same temperature and time. After that, the thickness change is measured. The thickness change and process time under different process parameters and environments are calculated to obtain the etching rate. The relevant data are collected and classified uniformly, so that the etching rate data of various process parameters, different materials under different formulations, different temperatures and different times can be obtained. Subsequently, these acquired etching rate data will be stored in a special database. With the continuous accumulation of experimental data, the content of the database will continue to increase, thus forming an etching rate database covering different material systems, liquid formulations, temperature, time and other parameters. The constructed etching rate database will include the etching rates of different materials. In different combinations of drug formulations and parameters, different materials will show different etching rates, which makes it easier for users to understand the precautions and key points to be avoided in process design. After having such a comprehensive and detailed etching rate database, some data analysis methods and algorithms can be used to deeply explore the trends reflected in the internal data of this database and conduct in-depth analysis of the data in the database. Assume that under the same liquid formula, the etching rate of material A is 10 nanometers / minute and the etching rate of material B is 1 nanometer / minute. By dividing the etching rates of different material combinations under the same liquid formula conditions, the difference in etching rates of the two materials under this formula can be intuitively compared, that is, the etching selectivity. In this example, the etching selectivity ratio of material A to material B is 10:1, which means that under this liquid formula, the etching speed of material A is 10 times that of material B. If the main goal is to etch material A and retain material B, and A and B are on different parts of the same device or product, this material A has an etching selectivity ratio of up to 10:1 for material B, which is obviously a good result. This value shows that material A can be etched well and has little effect on material B. For the material combination with high etching selectivity screened by the above method, the key parameters can be further optimized using machine learning methods, such as multivariate linear regression and artificial neural network.The input is material and drug information, such as silicon dioxide material in a solution mixed with phosphoric acid, nitric acid and acetic acid in a certain proportion. The model is modeled through a large amount of formula data with different proportions. Then the concentration and temperature of the key components in these solutions are extracted as key parameters. The output is the etching rate of silicon dioxide. The trained model can be used for subsequent formula prediction and optimization. For example, the etching rate of silicon dioxide is expected to be 8 nanometers / minute, and the etching rate obtained by the current formula is 6 nanometers / minute. Taking 8 nanometers / minute as the output target of the model, the optimization model can find out which parameters, such as the content and temperature of phosphoric acid, can increase the rate to 8 nanometers / minute without affecting other indicators, such as not corroding metals exposed to corrosive agents. The model can also provide an approximate parameter range or value, so that targeted etching rate optimization can be achieved in a relatively short time. Based on the key solution formula parameters and guiding optimization directions obtained by machine learning model analysis, new solutions are actually prepared and experiments are carried out. For example, if the model prompts to increase the content of phosphoric acid in the solution and increase the temperature, you can try to modify one or two parameters in a certain proportion. The optimized solution formula, such as increasing the phosphoric acid concentration to a certain percentage and setting the temperature to a certain degree Celsius, is then used for the above-mentioned material combination with high etching selectivity. Through experiments, the actual etching rate of each material under this new formula can be accurately determined. After etching, the change in film thickness is measured so that the etching rate can be accurately calculated. And check the etching surface morphology, etc. to determine whether the new parameters or ranges are reasonable. After obtaining these data, a comprehensive evaluation is conducted on the optimization results. For example, whether the etching rate of silicon dioxide is really increased to about 8 nanometers / minute as expected. If it is found that the optimization results meet expectations, then the solution formula has been successfully improved. On the contrary, if the actual effect does not meet the standard, or other negative effects are generated, such as corrosion of metals, it is necessary to return to the machine learning model analysis step, fit the newly acquired data through further learning, check the difference between the new experimental data and the previous ones, and then continue to conduct more in-depth parameter adjustment guidance or optimization target specification, improve the solution formula and conduct experimental verification again. Through the above means, after repeated iterations and optimizations, the optimal liquid medicine formula is finally obtained. Under the guidance of the optimal medicine formula, the etching process can be used in the production of integrated circuits or chips, improving processing efficiency and yield, and promoting the development of advanced processes.

[0022] S102. For the selected material combinations, first-principles calculations are used to study the interaction mechanism between each component in the liquid and the material surface, and the kinetic parameters of the surface reaction are obtained. Molecular dynamics simulation is performed to reveal the microscopic mechanism of the liquid diffusion and mass transfer process. A multi-scale coupled etching process model is established, including a molecular-scale surface reaction model, a mesoscopic-scale diffusion and mass transfer model, and a macroscopic-scale etching morphology evolution model. Through the coupling of these models, the rate-controlling steps and chemical reaction paths of material removal are clarified, guiding the optimization of liquid formulation.

[0023] S102 also includes: constructing a molecular model of the interaction between the material surface and the liquid medicine components for the screened material combination; using first-principles calculations to obtain the energy change and reaction rate constant kinetic parameters of the surface reaction in the molecular model; establishing a diffusion and mass transfer model of the liquid medicine near the material surface at a mesoscopic scale, and calculating the concentration distribution and mass transfer flux of the liquid medicine components; achieving a coordinated simulation of molecular-scale surface reactions, mesoscopic-scale diffusion and mass transfer, and macroscopic-scale etching morphology evolution through the coupling of multi-scale models; and revealing the rate-controlling steps and chemical reaction paths of the material removal process based on the coordinated simulation results.

[0024] Exemplary, according to the selected material combination, a molecular model of the interaction between the material surface and the liquid component is constructed, and the first principle calculation is used to obtain the kinetic parameters such as the energy change and reaction rate constant of the surface reaction. For example, for the interaction between silicon-based materials and hydrofluoric acid solution, the adsorption model of the silicon surface and the hydrofluoric acid molecule can be constructed, and the energy change of the adsorption energy, dissociation energy and reaction path of the hydrofluoric acid molecule on the silicon surface can be calculated by using the first principle calculation method such as density functional theory (DFT). Through these calculations, the activation energy and reaction rate constant of the silicon surface and the hydrofluoric acid reaction can be obtained, and these parameters are the basis for understanding the etching process. Molecular dynamics simulation is used to study the adsorption, desorption and diffusion behavior of each component in the liquid on the material surface, and the microscopic mechanism and diffusion coefficient of the diffusion mass transfer process are obtained. Taking silicon dioxide and phosphoric acid solution as an example, through molecular dynamics simulation, the adsorption and desorption process of phosphoric acid molecules on the silicon dioxide surface and its diffusion behavior on the surface can be observed. By statistically simulating the motion trajectory of molecules during the simulation, the diffusion coefficient of phosphoric acid on the surface of silicon dioxide can be calculated. These data help to understand the mass transfer mechanism of the liquid components on the surface of the material. Based on the surface reaction kinetic parameters and diffusion mass transfer parameters, a molecular-scale surface reaction model is established to describe the chemical reaction process and reactant concentration changes on the surface of the material. For example, based on the kinetic parameters of the reaction between silicon and hydrofluoric acid calculated above, a reaction model describing the change of hydrofluoric acid concentration on the silicon surface over time can be constructed. This model can simulate the rate of chemical reaction on the silicon surface and the consumption of reactants under different hydrofluoric acid concentrations. At the mesoscopic scale, considering the flow and mass transfer process of the liquid near the surface of the material, a diffusion mass transfer model is established to calculate the concentration distribution and mass transfer flux of the liquid components. Taking tungsten diselenide and nitric acid solution as an example, the flow field and concentration distribution of nitric acid solution near the surface of tungsten diselenide can be simulated by computational fluid dynamics (CFD) method. Through this simulation, the concentration change and mass transfer flux of nitric acid at different positions can be obtained, so as to understand the mass transfer process of the liquid on the surface of the material. At the macroscopic scale, according to the chemical reaction on the material surface and the mass transfer process of the liquid, an etching morphology evolution model is established to simulate the change of the material surface morphology over time and predict the etching depth and etching rate. For example, for the etching process of single crystal silicon and potassium hydroxide solution, a macroscopic model including chemical reaction and mass transfer process can be established to simulate the morphological changes on the surface of single crystal silicon at different time points and predict the etching depth and etching rate. Through the coupling of multi-scale models, the coordinated simulation of molecular-scale surface reaction, mesoscopic-scale diffusion mass transfer and macroscopic-scale etching morphology evolution is realized to reveal the rate-controlling steps and chemical reaction paths of the material removal process. Taking the etching of polycrystalline silicon and acidic solution as an example, by coupling the molecular-scale reaction model, the mesoscopic-scale mass transfer model and the macroscopic-scale morphology evolution model, the etching process of polycrystalline silicon in acidic solution can be systematically simulated to reveal the control steps and main chemical reaction paths of the etching rate.According to the coupled simulation results, the effects of process parameters such as the concentration of the chemical components and temperature on the etching rate and etching morphology are analyzed, the chemical formula and process conditions are optimized, and the selectivity and uniformity of material removal are improved. For example, through simulation, it is found that increasing the concentration of hydrofluoric acid can increase the etching rate of silicon, but too high a concentration will lead to uneven etching. By optimizing the concentration and temperature of hydrofluoric acid, the uniformity and selectivity of etching can be improved while ensuring a high etching rate. Through the above multi-scale simulation and analysis, we can deeply understand the microscopic mechanism of the interaction between materials and chemical solutions, guide the optimization of chemical formula and process conditions, and ultimately improve the efficiency of the etching process and product quality. This method is not only applicable to silicon-based materials in integrated circuit manufacturing, but can also be extended to the etching process research of other new materials, and has broad application prospects.

[0025] S103. Using microfluidic chip technology, design and integrate multiple parallel etching reaction chambers to achieve high-throughput screening of liquid medicine formulas. By controlling the flow rate of the injection pump and the liquid medicine ratio, a concentration gradient is formed in the chip. The composition of the liquid medicine at the chip outlet is monitored in real time by spectroscopy to obtain etching rate data under different formulas. A machine learning algorithm is used to establish a quantitative relationship between the formula and the etching performance, and preferably obtain a liquid medicine formula with high etching selectivity.

[0026] S103 also includes: obtaining the number of multiple parallel etching reaction chamber structures formed in the chip, and configuring the initial component ratios of the corresponding number of liquid medicines with component ratios at the inlets of different reaction chambers according to the number; using spectroscopy to detect the preset number of types of liquid medicines corresponding to the outputs of different reaction chambers at the outlet of the chip, and judging whether there is a liquid medicine component of a certain reaction chamber in the detected liquid medicines that meets the preset spectral characteristic peak threshold range, and if so, obtaining the position data of the etching reaction chamber corresponding to the liquid medicine on the chip; according to the position data, obtaining the liquid medicine concentration information and flow rate control information of the corresponding injection pump, and obtaining the preset number of liquid medicine component information data of different concentrations, and converting the liquid medicine component into the liquid medicine component. The information data is associated one by one with the serial numbers of the corresponding multiple reaction chambers on the chip; according to the different component ratios of the liquid medicine corresponding to the reaction chambers at different positions on the same chip, the performance data of different etching rates corresponding to the liquid medicines with different component ratios are determined through the association to obtain an etching performance data set; a random forest algorithm is used to analyze and process the obtained liquid medicine component data of different concentrations and the etching performance data to establish an etching performance prediction model; the liquid medicine etching formula is screened in the form of input variable optimization to determine whether the output variables obtained from the input formula data set of the etching performance prediction model meet the preset conditions, and if so, the corresponding liquid medicine formula is determined as the preferred liquid medicine formula.

[0027] Exemplarily, according to the different flow rate changes of the injection pump within the set time, different component ratios of the liquid at the inlet of the chip are configured. For example, assuming that an etching liquid containing hydrofluoric acid (HF) and nitric acid (HNO3) needs to be configured, the injection pump can accurately control the flow rate to deliver different volumes of HF and HNO3 within the set time, thereby realizing the configuration of liquids with different component ratios. Assuming that the set time is 10 minutes, the injection pump delivers HF at a flow rate of 10 ml per minute in the first 5 minutes, and delivers HNO3 at a flow rate of 5 ml per minute in the last 5 minutes, and finally obtains a liquid with a volume ratio of HF and HNO3 of 2:1. According to the number of multiple parallel etching reaction chamber structures formed in the chip, the initial component ratios of the liquids with corresponding number of component ratios at the inlets of different reaction chambers are configured. Assuming that there are four parallel etching reaction chambers on the chip, four liquids with different component ratios can be configured respectively. For example, reaction chamber 1 is configured with a liquid medicine with a volume ratio of HF and HNO3 of 2:1, reaction chamber 2 is configured with a liquid medicine with a volume ratio of 3:1, reaction chamber 3 is configured with a liquid medicine with a volume ratio of 1:1, and reaction chamber 4 is configured with a liquid medicine with a volume ratio of 4:1. Spectroscopy is used to detect the various liquid medicines outputted from different reaction chambers at the chip outlet. Spectroscopy can determine the composition and concentration of the liquid medicine by measuring the spectral characteristic peaks of the liquid medicine. For example, a UV-visible spectrometer is used to detect the liquid medicine at the reaction chamber outlet. If a characteristic absorption peak of HF is detected at a specific wavelength in the liquid medicine of a certain reaction chamber, and the peak intensity is within a preset threshold range, it means that the liquid medicine composition of the reaction chamber meets expectations. The position data of the etching reaction chamber corresponding to the liquid medicine on the chip is obtained. Assuming that the liquid medicine composition of reaction chamber 2 meets the preset threshold range of the spectral characteristic peak, the specific position information of reaction chamber 2 on the chip is recorded, such as the coordinates (2,3). The liquid medicine concentration information and flow rate control information of the injection pump corresponding to the position data of the reaction chamber on the chip are obtained. According to the recorded position information, the control log of the injection pump is consulted to know that the liquid configuration corresponding to the reaction chamber 2 is HF and HNO3 with a volume ratio of 3:1. The injection pump delivers HF at a flow rate of 10 ml per minute in the first 6 minutes and delivers HNO3 at a flow rate of 5 ml per minute in the last 4 minutes. The composition information data of the liquid medicine of various concentrations are obtained, and the obtained information data are associated with the corresponding numbers of the multiple reaction chambers on the chip. For example, the liquid medicine composition information of reaction chambers 1, 2, 3, and 4 is recorded as (2:1, 3:1, 1:1, 4:1) respectively, and corresponds to their respective numbers (1, 2, 3, 4) one by one. According to the liquid medicines with different composition ratios corresponding to the reaction chambers at different positions on the same chip, the performance data of different etching rates corresponding to the liquid medicines with different composition ratios are determined through data association. For example, through experimental measurement, it is found that the etching rate of reaction chamber 1 is 5 microns / minute, that of reaction chamber 2 is 7 microns / minute, that of reaction chamber 3 is 4 microns / minute, and that of reaction chamber 4 is 8 microns / minute. These etching rate data are associated with the corresponding chemical solution component ratios to form an etching performance data set.The random forest algorithm is used to analyze and process the obtained liquid composition data and etching performance data of different concentrations. The random forest algorithm is a powerful machine learning algorithm that can process data with nonlinear relationships. By inputting the liquid composition ratio and etching rate data, the random forest algorithm can establish an etching performance prediction model. The liquid etching formula is screened by optimizing the input variables. For example, the liquid data with different composition ratios are input into the prediction model, and the model outputs the corresponding etching rate prediction value. If the predicted value meets the set conditions, such as the etching rate is greater than 6 microns / minute and the uniformity is good, the liquid formula is considered to be the preferred formula. When the output variables obtained from the input formula data set of the etching performance prediction model meet the set conditions, the preferred liquid formula is obtained and the screening process is terminated. For example, after multiple simulations and optimizations, it is found that the liquid with a volume ratio of HF and HNO3 of 3:1 shows the best etching performance in the prediction model, meets the set conditions, and finally determines that the formula is the preferred liquid formula. Through the above method, not only can the optimal liquid formula be efficiently screened, but also the influence of different component ratios on etching performance can be deeply understood, guiding the process optimization in actual production. The application of this method can significantly improve the efficiency of the etching process and the product quality, and has broad application prospects.

[0028] S104. Based on the screening results of the microfluidic chip, design a spray head array and a stirring device for the preferred liquid medicine formula to form a uniform liquid medicine spray and turbulent flow field. By optimizing the nozzle shape, number of nozzles, spray angle and stirring blade parameters, the mass transfer process between the liquid medicine and the material surface is strengthened to improve the etching uniformity. The liquid medicine recovery device adopts temperature and conductivity dual parameter control, and adjusts the recovery time in real time according to the change of liquid medicine properties to reduce liquid medicine waste.

[0029] S104 also includes: constructing a three-dimensional model of the channel according to the channel data of the microfluidic chip, and gridding the three-dimensional model; obtaining the conductivity at different positions inside the flow channel, if there are differences in the conductivity curves at different positions, calculating the standard deviation between the values ​​at different positions, if the standard deviation is less than a preset threshold, calculating the average value of the slope of the grayscale value change of each column of pixels, obtaining the impeller rotation angle control value, and adjusting the impeller angle according to the control value; obtaining the adjusted spray parameters of the spray array, if the spray flow curve variation range value is lower than the preset value, obtaining the flow value, conductivity value and temperature value at this time, calculating and digitizing the turbulent field streamline curvature information, and judging whether the flow value, conductivity value, temperature value and turbulent field streamline curvature value meet the preset requirements.

[0030] Exemplarily, according to the channel data of the microfluidic chip, a three-dimensional model of the corresponding size channel is constructed using COMSOL Multiphysics software, and the model is meshed. For example, assuming that the channel width of the microfluidic chip is 200 microns, the depth is 50 microns, and the length is 1 cm, first create a three-dimensional geometric model in COMSOL and accurately set these size parameters. Subsequently, the model is meshed to ensure that the mesh density is moderate, which can not only ensure the calculation accuracy but also improve the calculation efficiency. The velocity of the fluid inside the three-dimensional model is obtained to generate a three-dimensional velocity field map. The distance h between the nozzle outlet and the material surface and the characteristic length L of the material surface are calculated to obtain the Reynolds number Re. Assuming that the fluid is water, the density ρ is 1000 kg / m3, the characteristic velocity v is 0.1 m / s, and the dynamic viscosity μ is 0.001 Pa·s, then the Reynolds number Re=ρvL / μ=1000×0.1×0.01 / 0.001=1000. According to the calculated Reynolds number, the Reynolds number range is determined and the appropriate turbulence model is selected. If the Reynolds number is less than 2300, the laminar flow model is selected; if it is between 2300 and 4000, the transitional flow model is selected; if it is greater than 4000, the turbulence model is selected. When the blade speed parameter of the stirring device in the COMSOL simulation turbulent field is the same as the actual set blade speed parameter, and the turbulent field distribution data meets expectations, the three-dimensional model is judged to meet the standard. For example, the blade speed is set to 1000 rpm, and the simulation results show that the turbulent field is evenly distributed and the flow velocity distribution meets expectations, then the model is considered to meet the standard. Then, combined with the channel material parameters of the microfluidic chip, the chip entity material properties are assigned to the three-dimensional model. Assuming that the chip material is polydimethylsiloxane (PDMS), its physical properties such as elastic modulus, Poisson's ratio and other parameters are input into the model. After the simulation operation, the flow value, conductivity, and pressure value at different positions are obtained, and these data are used as data labels for the deep learning algorithm. After the preferred liquid is sprayed into the flow channel, images of the spraying process are collected according to the preset time interval to form spray images at different times. For example, an image is collected every 1 second for 10 minutes to obtain 600 frames of images. According to the obtained continuous frame images, each frame of the image is split into sub-images by column, and each column of pixels is defined as a sub-image. The gray value distribution information of these sub-images is used as data features and as data input to the deep learning neural network for training. If the real-time data collected at multiple points inside the flow channel differs from the labeled data by more than 5% after being output by the algorithm, it means that the data does not meet the requirements. After failing to meet the requirements, the convolutional neural network is used to analyze the pixel average value change data of each column of sub-images and compare it with the existing images in the database. If the similarity is less than the set value of 65%, it is judged that a new working condition has occurred, the liquid composition ratio is adjusted, and the new liquid is re-optimized. At the same time, the optimization algorithm is used to change the parameters of the sprinkler array to generate spray images with different spray angles.For example, if the current liquid medicine is HF and HNO3 with a volume ratio of 3:1 and the similarity is only 60%, then the ratio is adjusted to 4:1 and the experiment is repeated. Data is collected under the new working condition and input into the database as a data set to expand the data set. If the similarity is greater than 65%, the spray parameters of the sprinkler head, such as the spray angle and spray pressure, are adjusted. According to the temperature values ​​at different positions inside the flow channel measured in real time, if the variance of the temperature values ​​collected twice before and after exceeds 20%, the liquid outlet aperture of the sprinkler head is changed. For example, if the temperature is measured at 25°C for the first time and 35°C for the second time, and the variance is 100, the liquid outlet aperture is adjusted from 0.5 mm to 0.6 mm. According to the conductivity at different positions inside the flow channel measured in real time, if there is a difference in the conductivity curves at different positions, the standard deviation between the values ​​at different positions is calculated. If the calculated standard deviation is less than the preset threshold, the impeller rotation angle control value is obtained by calculating the average slope of the gray value change of each column of pixels, and the impeller angle is controlled by the control value. For example, if the standard deviation is 0.02, which is less than the preset threshold value of 0.05, the average value of the gray value change slope is calculated to be 0.1, and the impeller rotation angle is controlled to be 10 degrees, so that the streamline curvature of the turbulent field formed by the blades is consistent with the undulation of the material surface. The adjusted spray array spray parameters are obtained. When the spray flow curve change range value is less than 1%, the flow value, conductivity value, and temperature value are obtained at this time. At the same time, the turbulent field streamline curvature information obtained by real-time measurement is calculated and digitized to determine whether the flow, conductivity, temperature, and turbulent field streamline curvature values ​​meet the requirements. For example, the flow change range is 0.5%, the conductivity is 1.2 Siemens / meter, the temperature is 30°C, and the streamline curvature is 0.15, all of which meet the preset requirements. The difference between the conductivity, temperature and other parameter values ​​at the end of the spray and the start of the spray is calculated, and the obtained data group is used to characterize the properties of the liquid at this time. Assuming that the conductivity is 1.0 Siemens / m at the beginning of spraying and 1.2 Siemens / m at the end, with a difference of 0.2, the solution information update value is obtained in combination with the temperature change, and the control parameters of the liquid medicine recovery device are adjusted using the updated value. According to the feedback information of the recovery effect of the liquid medicine recovery device after the parameters are changed, the relationship model of temperature, conductivity and liquid medicine property parameters is corrected to achieve the optimized operation of the liquid medicine recovery system. Through the above method, not only can the fluid behavior in the microfluidic chip be accurately controlled, but also the spray parameters and liquid medicine composition can be adjusted in real time to ensure the stability and efficiency of the etching process. This technical means of comprehensive use of simulation, image processing and machine learning has significantly improved the application effect of microfluidic chips in the etching process and provided strong support for high-precision manufacturing.

[0031] S105. Install a multi-parameter online monitoring unit in the etching equipment, use a fiber optic spectrometer to measure the concentration of the liquid in real time, and use a thermocouple and a pH electrode to measure the temperature and pH of the liquid in real time. When the monitoring parameters exceed the set range, start the liquid ratio adjustment program, and realize dynamic regulation of the liquid properties by controlling the liquid inlet and discharge volume to ensure the stability and consistency of the etching process.

[0032] S105 also includes: obtaining real-time spectral data in the drug liquid pool, determining the drug liquid concentration data according to a pre-established spectral database, and if the difference between the measured drug liquid concentration and the target concentration is greater than a preset threshold, calculating the concentration difference, starting the drug liquid concentration adjustment algorithm, and generating control data for controlling the inlet and outlet of the drug; establishing a recurrent neural network model for different proportions of multi-component drug liquids, inputting data including drug liquid concentration, temperature, pH, inlet volume and outlet volume at historical moments, and a pre-established drug liquid ratio concentration curve, predicting the process parameter values ​​at future moments, and obtaining the inlet valve and outlet valve opening adjustment strategy; if the error between the predicted result and the target set value is greater than a preset threshold, performing model adaptation through an artificial neural network, collecting real-time data to fit the historical scheme for training, obtaining model adjustment data, and determining new process parameter adjustment weights.

[0033] Exemplarily, the real-time spectral data in the liquid medicine pool is collected by the fiber optic spectrometer in order to accurately monitor the real-time changes of the liquid medicine components. For example, assuming that the main components in the liquid medicine pool are sulfuric acid and hydrochloric acid, the fiber optic spectrometer can capture the characteristic spectra of these two acids in real time. Through the pre-established spectral database, these characteristic spectra can be compared with the standard spectra, and the liquid medicine concentration data of the corresponding components in the real-time spectral data are determined using chemometric methods such as partial least squares (PLS). Assuming that the target sulfuric acid concentration is 5000mol / L, and the measured concentration is 4999mol / L, the difference is 1mol / L, which is greater than the preset 0.001mol / L threshold. At this time, the system will start the liquid medicine concentration adjustment algorithm to generate control data for controlling liquid inlet and liquid discharge. These control data are calculated by the concentration difference data and the real-time liquid inlet speed, and then the size of the liquid inlet valve and the liquid discharge valve is adjusted to ensure that the liquid medicine concentration quickly returns to the target value. The size adjustment of the liquid inlet valve and the liquid discharge valve depends not only on the concentration difference, but also on the real-time liquid medicine temperature data and historical data. For example, when constructing the thermodynamic equation of real-time liquid inflow rate, the Fourier series model is used for fitting, and the number of series is set to 50. Assuming the target temperature is 25°C, the fitting coefficients of the Fourier series equation will change at this temperature. By adjusting these coefficients, the liquid inflow volume at this time can be accurately determined. This method can effectively deal with the impact of temperature changes on the fluidity of the drug solution and ensure the precise control of the liquid inflow volume. In terms of pH control, the target pH value and temperature data are key parameters. Assuming the target pH value is 8.00, the measured pH value is 7.00, the temperature is 20°C, and the concentration is 5000 mol / L, the pH difference is 1.00 at this time. Using the integral control method of step-by-step liquid inflow, the liquid inflow time can be accurately determined, and the liquid inflow control accuracy reaches 0.001 mL. This method gradually approaches the target pH value through small step adjustments, avoids fluctuations caused by large adjustments, and improves the stability of control. The proportioning model uses a support vector machine regression algorithm, and the input is the concentration, temperature and pH data in the historical data. Assuming that the historical database contains 1,000 sets of experimental measurement data, the preset regression model generated by these data can output the inflow and outflow of different liquid medicines. For example, based on the current state of the liquid medicine pool, the model predicts that 100 mL of sulfuric acid and 50 mL of hydrochloric acid need to be added to ensure that the liquid medicine composition reaches the preset ratio. This method uses a large amount of historical data to improve the accuracy of the prediction and reduce the trial and error cost in the experiment. The purpose of modeling the process parameter time series data using a recurrent neural network is to predict the trend of future process parameter changes. Assuming that the input includes liquid medicine concentration, temperature, pH, inflow and outflow data in the past 100 moments, the pre-established liquid medicine ratio concentration curve can help predict the process parameter values ​​in the next 30 moments.For example, based on the current data, the model predicts that the concentration of the liquid will rise by 0.5% in the next 10 minutes, and the opening of the liquid inlet valve and the liquid discharge valve needs to be adjusted accordingly to maintain the stability of the liquid composition. This method predicts the changes in process parameters in advance through time series analysis and realizes forward-looking control. If the error between the liquid concentration, temperature and pH in the prediction results of the recurrent neural network and the target set value is greater than 5%, the system will learn the model adaptively through the artificial neural network. For example, collect the data of the first 10,000 groups of real-time liquid inlet speed, liquid inlet volume, and pH measured values ​​on the production line, fit the historical liquid inlet scheme and obtain the model adjustment data. Assuming that the training results show that the liquid inlet speed needs to be increased by 10% and the pH value needs to be adjusted by 0.2 units, these adjustment weights will be input into the liquid inlet controller to ensure that the new process parameters are more accurate. According to the adjustment weight, update the weight coefficient inside the control logic judgment algorithm. Assuming that the difference between the new weight coefficient and the original weight coefficient is greater than 1, the system will perform real-time valve pressure monitoring to ensure the stability of the valve pressure. For example, the pressure of the inlet valve and the outlet valve is monitored in real time by the pressure sensor. If the pressure fluctuation exceeds the preset range, the system will immediately adjust the valve opening to ensure the stability of the flow of the liquid medicine. This method further improves the reliability and stability of the control through real-time monitoring and adjustment. Through the above series of measures, not only the precise control of the liquid medicine composition is achieved, but also the stability and efficiency of the entire process are improved. The real-time monitoring of the fiber optic spectrometer, the precise analysis of chemometrics, the fitting of thermodynamic equations, the application of the support vector machine regression algorithm, the prediction of the recurrent neural network, and the adaptive adjustment of the artificial neural network jointly construct a multi-level and multi-dimensional control system, which ensures the stability of the liquid medicine composition and the accuracy of the process parameters, and provides strong support for high-precision manufacturing.

[0034] S106. Characterize and evaluate the etching results after dynamic regulation. Using etching depth, surface roughness and etching residue as evaluation indicators, optical microscopy and atomic force microscopy are used to characterize the etching results. By calculating the uniformity index and selectivity index, the etching performance of different material systems and solution formulations is quantitatively evaluated. The uniformity index calculation formula is the ratio of the standard deviation of the etching depth to the average value, and the selectivity index calculation formula is the ratio of the etching rates of the two materials. For the performance evaluation results, gray correlation analysis is used to determine the key process parameters to guide the learning and iterative update of the intelligent optimization system.

[0035] S106 also includes: obtaining the three-dimensional morphology data of the sample after etching, scanning the sample surface respectively with an optical microscope and an atomic force microscope, and obtaining a height data matrix. According to the height data matrix, the etching depth, surface roughness and etching residue evaluation index are calculated. Among them, the etching depth is determined by the height difference between the reference plane and the etching surface, the surface roughness is characterized by the arithmetic average roughness Ra, and the etching residue is obtained by setting a threshold to judge the discrete degree of the height data. For samples of different material systems, the etching rate is measured. The liquid formula is used as input and the etching rate is used as output to establish a mapping relationship between the liquid formula and the etching rate. When the liquid formula changes, the corresponding etching rate can be quickly predicted. Calculate the uniformity index and the selectivity index. The uniformity index is obtained by the ratio of the standard deviation of the etching depth to the average value, reflecting the uniformity of the etching depth. The selectivity index is obtained by the ratio of the etching rates of the two materials, reflecting the difference in the etching ability of the liquid formula for different materials. Grey correlation analysis is used to determine the key process parameters affecting the etching performance. The uniformity index and selectivity index are taken as parent factors, and the liquid formula and process parameters are taken as sub-factors. The grey correlation between each sub-factor and the parent factor is calculated. The greater the correlation, the greater the influence of the factor on the etching performance. The results of the grey correlation analysis are used to guide the learning of the intelligent optimization system. The uniformity index and selectivity index are taken as optimization targets, the liquid formula and key process parameters are taken as optimization variables, and the genetic algorithm is used to search for the optimal solution. When the intelligent optimization system completes an iteration, the experimental results are introduced into the sample set to update the knowledge base of the intelligent optimization system. The above steps are repeated continuously to realize the autonomous learning and iterative update of the intelligent optimization system. As the optimization iteration proceeds, the intelligent optimization system can summarize the rules from a large amount of experimental data, continuously improve the prediction accuracy and optimization efficiency, and finally obtain the optimal liquid formula and process parameter combination that takes into account uniformity and selectivity.

[0036] Exemplarily, obtaining three-dimensional morphological data of the sample after etching is an important part of the etching process evaluation. First, the sample surface is scanned respectively using an optical microscope and an atomic force microscope. The optical microscope can quickly obtain a large range of surface morphology through the principle of optical imaging, while the atomic force microscope provides more accurate three-dimensional height data through the interaction between the probe and the sample surface. For example, the height data matrix obtained from a certain scan shows that the highest point of a certain area on the sample surface is 5 microns and the lowest point is 1 micron, forming a three-dimensional morphology map with a height difference of 4 microns. According to the height data matrix, the etching depth, surface roughness and etching residue evaluation indicators can be calculated. The etching depth is determined by the height difference between the reference plane and the etching surface. Assuming that the height of the reference plane is 2 microns and the height of the lowest point of the etching surface is 1 micron, the etching depth is 1 micron. The surface roughness is characterized by the arithmetic mean roughness Ra. Assuming that the heights of each point in the height data matrix are 2.1, 2.2, 2.0, and 2.3 microns, respectively, Ra is calculated as (2.1+2.2+2.0+2.3) / 4=2.15 microns. The etching residue is determined by setting a threshold to determine the discreteness of the height data. For example, the threshold is set to 0.5 microns. If more than 10% of the points in the height data differ from the average height by more than 0.5 microns, it is determined that there is etching residue. For samples of different material systems, their etching rates are measured. For example, a certain liquid formulation has an etching rate of 10 microns / minute for silicon material and an etching rate of 5 microns / minute for silicon oxide. The liquid formulation is used as input and the etching rate is used as output to establish a mapping relationship between the liquid formulation and the etching rate. When the liquid formulation changes, such as adjusting the concentration of a certain chemical component, the corresponding etching rate can be quickly predicted through the mapping relationship. Calculating the uniformity index and the selectivity index are important indicators for evaluating etching performance. The uniformity index is obtained by the ratio of the standard deviation of the etching depth to the average value, reflecting the uniformity of the etching depth. Assuming that in a certain experiment, the average value of the etching depth is 3 microns and the standard deviation is 0.5 microns, the uniformity index is 0.5 / 3≈0.167. The selectivity index is obtained by the ratio of the etching rates of the two materials, reflecting the difference in the etching ability of the solution on different materials. For example, the etching rates of silicon and silicon oxide are 10 microns / minute and 5 microns / minute, respectively, and the selectivity index is 10 / 5=2. Grey correlation analysis is used to determine the key process parameters affecting the etching performance. The uniformity index and selectivity index are taken as parent factors, and the concentration of each component, temperature, and pH value in the solution formula are taken as sub-factors, and the grey correlation between each sub-factor and the parent factor is calculated. For example, it is found through calculation that the grey correlation of hydrofluoric acid concentration is 0.8, which is significantly higher than other factors, indicating that the concentration of hydrofluoric acid has a greater impact on the etching performance. With the uniformity index and selectivity index as the optimization targets, the genetic algorithm is used for intelligent optimization. The initial population contains a variety of drug solution formulas and process parameter combinations, which are continuously iterated and optimized through operations such as crossover and mutation.For example, after one iteration, a set of optimal parameter combinations is obtained: hydrofluoric acid concentration of 10%, temperature of 30°C, pH value of 4, corresponding to a uniformity index of 0.1 and a selectivity index of 3. The experimental results are introduced into the sample set to update the knowledge base of the intelligent optimization system. For example, the newly introduced experimental data show that when the hydrofluoric acid concentration is 12%, the uniformity index is significantly improved to 0.08. The system adjusts the optimization strategy accordingly and increases the weight of the high-concentration hydrofluoric acid formula. As the optimization iteration proceeds, the intelligent optimization system summarizes the rules from a large amount of experimental data and continuously improves the prediction accuracy and optimization efficiency. For example, after 100 iterations, the error between the etching rate predicted by the system and the actual value dropped from 5% to 1%, significantly improving the optimization effect of the process parameters. Through the above steps, the intelligent optimization system can continuously learn and optimize, and finally obtain the optimal solution formula and process parameter combination that takes into account uniformity and selectivity, significantly improving the stability and efficiency of the etching process.

[0037] S107. Based on the process optimization experimental data, a nonlinear mapping model between the solution formula, process parameters and etching results is constructed. The model adopts a multi-layer neural network structure, with the input layer being the solution formula and process parameters, and the output layer being the etching depth, uniformity and selectivity indicators. The obtained etching rules and control strategies are formed into dynamically updated knowledge reasoning rules. An intelligent decision-making system based on case reasoning and fuzzy logic is developed to realize the automation and intelligence of solution formula design, process parameter selection and etching defect diagnosis.

[0038] The S107 also includes: constructing a nonlinear mapping model between the liquid formula, process parameters and etching results, wherein the liquid formula and process parameters are used as input layers, and the etching depth, uniformity and selectivity indicators are used as output layers. For new liquid formula and process parameter combinations, the most similar formula design and parameter selection schemes are retrieved from historical optimization cases through case reasoning technology. An intelligent decision-making system is developed to realize the automation and intelligence of liquid formula design, process parameter selection and etching defect diagnosis by combining case reasoning and fuzzy logic.

[0039] Exemplarily, according to the process optimization experimental data, a multi-layer neural network is used to construct a nonlinear mapping model between the liquid formula, process parameters and etching results. Specifically, the input layer includes the concentration of each component in the liquid formula (such as hydrofluoric acid, nitric acid, etc.), process parameters (such as temperature, pH value, etching time, etc.), and the output layer is the key indicators such as etching depth, uniformity index and selectivity index. For example, in a certain experiment, the data of the input layer is 10% hydrofluoric acid concentration, 5% nitric acid concentration, temperature 30°C, pH value 4, etching time 30 minutes, and the results of the output layer are etching depth 3 microns, uniformity index 0.1, and selectivity index 2. By training and optimizing the constructed nonlinear mapping model, the etching law and control strategy between the liquid formula, process parameters and etching results are obtained. During the training process, a large amount of historical experimental data is used, and the network weights are continuously adjusted through the back propagation algorithm to make the model prediction results as close as possible to the actual experimental data. For example, after 1000 iterations of training, the error between the model predicted etching depth and the actual value is reduced from 10% to 2%, which significantly improves the prediction accuracy. The trained model is converted into dynamically updated knowledge reasoning rules. For example, the model finds that the concentration of hydrofluoric acid is positively correlated with the etching depth, while an increase in temperature will cause the uniformity index to decrease. These laws are converted into rules, such as "for every 1% increase in hydrofluoric acid concentration, the etching depth increases by 0.2 microns" and "for every 5°C increase in temperature, the uniformity index decreases by 0.05". Based on the acquired knowledge reasoning rules, case reasoning technology is used to perform similarity matching and retrieval on new liquid formulas and process parameter combinations based on historical optimization cases. For example, the newly proposed liquid formula is a hydrofluoric acid concentration of 12% and a nitric acid concentration of 6%. By calculating its similarity with historical cases, the system finds that it is highly similar to a successful case. The process parameters of this case are a temperature of 28°C and a pH value of 3.5, and the final etching effect is excellent. Based on this, the system recommends similar parameter combinations and is expected to obtain good etching results. In view of the uncertainty and fuzziness in the process of liquid formula design and process parameter selection, fuzzy logic theory is introduced to establish a fuzzy rule base. For example, the fuzzy rule base contains the rule "If the hydrofluoric acid concentration is high and the temperature is moderate, the etching depth is large and the uniformity is good". In specific operations, the parameters such as hydrofluoric acid concentration and temperature are fuzzy processed, such as "high hydrofluoric acid concentration" is defined as a concentration greater than 10%, and "moderate temperature" is defined as 25°C to 35°C. Through fuzzy reasoning, the system can give reasonable formula design and parameter selection suggestions when the parameters are not completely determined. Combining case reasoning and fuzzy logic, an intelligent decision-making system is developed to automate and intelligently process the solution formula design, process parameter selection and etching defect diagnosis. For example, after receiving a new etching task, the system first finds similar historical cases through case reasoning, and then uses fuzzy logic to fine-tune the parameters in the case to generate the optimal solution formula and process parameter combination.At the same time, the system can predict potential etching defects, such as uneven etching and excessive residue, based on historical data, and provide corresponding solutions. In the intelligent decision-making system, a human-computer interaction interface is set up to allow process experts to evaluate and modify the formula design and parameter selection scheme generated by the system. For example, the expert found that the hydrofluoric acid concentration recommended by the system was too high, which might cause equipment corrosion, so the concentration was adjusted to 9%, and the optimization result was fed back to the system. The system updates the knowledge base accordingly, and will give priority to recommending the adjusted concentration in similar situations in the future. Through continuous iterative optimization, the intelligent decision-making system can automatically generate the optimal solution formula and process parameter combination according to process requirements and target performance. For example, after one iteration, the formula generated by the system is 11% hydrofluoric acid concentration, 5.5% nitric acid concentration, 29°C temperature, pH value 3.8, etching depth of 3.5 microns, uniformity index 0.08, selectivity index 2.5, which significantly improves the stability and efficiency of the etching process. The application of this intelligent decision-making system not only improves decision-making efficiency and accuracy, but also reduces the number and cost of experiments. Through the optimization scheme generated by the system, process personnel can quickly find the best process parameters, avoiding a lot of trial and error experiments. At the same time, the continuous learning and updating of the system makes the process optimization process more intelligent and efficient, providing strong support for the continuous improvement of the etching process.

[0040] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent wafer etching process, characterized in that: The process includes the following steps: S101, obtaining etching rates of different materials under various chemical solution formulations, constructing an etching rate database covering various material systems, screening out material combinations with high etching selectivity by comparing etching rate differences in the database, and determining a targeted chemical solution formulation optimization direction; S102, for the selected material combinations, use first-principles calculations to study the interaction mechanism between each component in the solution and the material surface, obtain the kinetic parameters of the surface reaction, perform molecular dynamics simulation, reveal the microscopic mechanism of the diffusion and mass transfer process of the solution, and establish a multi-scale coupled etching process model, including a molecular-scale surface reaction model, a mesoscopic-scale diffusion and mass transfer model, and a macroscopic-scale etching morphology evolution model. Through the coupling of these models, the rate-controlling steps and chemical reaction paths of material removal are clarified, guiding the optimization of the solution formulation; S103, using microfluidic chip technology, designing and integrating multiple parallel etching reaction chambers to achieve high-throughput screening of drug solution formulas, forming a concentration gradient in the chip by controlling the flow rate of the sampling pump and the drug solution ratio, using spectroscopy to monitor the composition of the drug solution at the chip outlet in real time, obtaining etching rate data under different formulas, and using machine learning algorithms to establish a quantitative relationship between formula and etching performance, and preferably obtaining a drug solution formula with high etching selectivity; S104, based on the screening results of the microfluidic chip, a spray head array and a stirring device are designed for the preferred liquid medicine formula to form a uniform liquid medicine spray and turbulent flow field. By optimizing the nozzle shape, the number of nozzles, the spray angle and the stirring blade parameters, the mass transfer process between the liquid medicine and the material surface is strengthened to improve the etching uniformity. The liquid medicine recovery device adopts temperature and conductivity dual parameter control to adjust the recovery time in real time according to the change of liquid medicine properties to reduce liquid medicine waste; S105, installing a multi-parameter online monitoring unit in the etching equipment, using a fiber optic spectrometer to measure the concentration of the liquid in real time, using a thermocouple and a pH electrode to measure the temperature and pH of the liquid in real time, and when the monitoring parameters exceed the set range, starting the liquid ratio adjustment program, and realizing dynamic regulation of the liquid properties by controlling the liquid inlet and liquid discharge, to ensure the stability and consistency of the etching process; S106, characterize and evaluate the etching results after dynamic control, using etching depth, surface roughness and etching residue as evaluation indicators, using optical microscopy and atomic force microscopy to characterize the etching results, and quantitatively evaluate the etching performance of different material systems and liquid formulations by calculating uniformity index and selectivity index. The uniformity index calculation formula is the ratio of the standard deviation of the etching depth to the average value, and the selectivity index calculation formula is the ratio of the etching rates of the two materials. For the performance evaluation results, gray correlation analysis is used to determine the key process parameters to guide the learning and iterative update of the intelligent optimization system; S107, based on the process optimization experimental data, construct a nonlinear mapping model between the solution formula, process parameters and etching results. The model adopts a multi-layer neural network structure, with the input layer being the solution formula and process parameters, and the output layer being the etching depth, uniformity and selectivity indicators. The obtained etching laws and control strategies are transformed into dynamically updated knowledge reasoning rules, and an intelligent decision-making system based on case reasoning and fuzzy logic is developed to realize the automation and intelligence of solution formula design, process parameter selection and etching defect diagnosis.

2. The intelligent wafer etching process according to claim 1, characterized in that: The S101 also includes: Acquire etching rate data covering a preset number of types of material systems and liquid formulations, and establish the etching rate database; Using a data analysis algorithm to analyze the etching rate data in the etching rate database, and calculating the etching rate differences of different material combinations under the same solution formula; Determining the selectivity of the material combination according to the etching rate difference, and screening out the material combination with high etching selectivity; For the material combination with high etching selectivity, a machine learning algorithm is used to analyze the relationship between the material combination and the liquid medicine formula parameters to obtain the key liquid medicine formula parameters that affect the etching selectivity; Determining a drug solution formula optimization direction according to the key drug solution formula parameters, and optimizing the drug solution formula by adjusting the value range of the key drug solution formula parameters; The optimized solution formula is applied to the material combination with high etching selectivity, and the optimization effect is verified by etching experiments. If the preset optimization target is not achieved, the process returns to the solution formula optimization step and continues to optimize until the optimal solution formula is obtained to achieve high selective etching of the material.

3. The intelligent wafer etching process according to claim 1, characterized in that: The S103 also includes: Obtaining the number of multiple parallel etching reaction chamber structures formed in the chip, and configuring the initial component ratios of the corresponding number of drug solutions at the entrances of different reaction chambers according to the number; Detecting a preset number of types of liquid medicines outputted from different reaction chambers at the chip outlet by spectroscopy, determining whether the liquid medicine components of a certain reaction chamber among the detected liquid medicines meet the preset spectral characteristic peak threshold range, and if so, obtaining the position data of the etching reaction chamber corresponding to the liquid medicine on the chip; According to the position data, obtain the drug concentration information and flow rate control information of the corresponding injection pump, obtain a preset number of drug composition information data of different types and concentrations, and associate the drug composition information data with the corresponding numbers of the multiple reaction chambers on the chip one by one; According to the chemical solutions with different component ratios corresponding to the reaction chambers at different positions on the same chip, the performance data of different etching rates corresponding to the chemical solutions with different component ratios are determined by the association to obtain an etching performance data set; A random forest algorithm is used to analyze and process the obtained liquid composition data of different concentrations and the etching performance data, and an etching performance prediction model is established; The liquid etching recipe is screened by optimizing the input variables to determine whether the output variables obtained from the input recipe data set of the etching performance prediction model meet the preset conditions. If so, the corresponding liquid recipe is determined as the preferred liquid recipe.

4. The intelligent wafer etching process according to claim 1, characterized in that: The S104 also includes: According to the channel data of the microfluidic chip, a three-dimensional model of the channel is constructed, and the three-dimensional model is gridded; Obtaining the conductivity at different positions inside the flow channel, if there are differences in the conductivity curves at different positions, calculating the standard deviation between the values ​​at different positions, if the standard deviation is less than a preset threshold, calculating the average value of the slope of the gray value change of each column of pixels, obtaining the impeller rotation angle control value, and adjusting the impeller angle according to the control value; Get the adjusted spray parameters of the spray array. If the spray flow curve variation range is lower than the preset value, get the flow value, conductivity value and temperature value at this time, calculate and digitize the turbulent field streamline curvature information, and judge whether the flow value, conductivity value, temperature value and turbulent field streamline curvature value meet the preset requirements.

5. An intelligent wafer etching process according to any one of claims 1 to 4, characterized in that: The S105 also includes: Obtain real-time spectral data in the drug liquid pool, determine the drug liquid concentration data based on the pre-established spectral database, and if the difference between the measured drug liquid concentration and the target concentration is greater than the preset threshold, calculate the concentration difference, start the drug liquid concentration adjustment algorithm, and generate control data for controlling liquid inlet and outlet; According to the different proportions of multi-component liquid medicine, a recurrent neural network model is established. The input includes the liquid concentration, temperature, pH, liquid inlet and liquid discharge data at historical moments, as well as the pre-established liquid ratio concentration curve, to predict the process parameter values ​​at future moments and obtain the opening adjustment strategy of the liquid inlet valve and liquid discharge valve. If the error between the predicted result and the target setting value is greater than the preset threshold, the model is adapted through the artificial neural network, real-time data is collected to fit the historical plan for training, the model adjustment data is obtained, and the new process parameter adjustment weights are determined.

6. An intelligent wafer etching process according to any one of claims 1 to 4, characterized in that: The S106 also includes: Obtain the three-dimensional morphology data of the sample after etching, and use optical microscope and atomic force microscope to scan the sample surface respectively to obtain the height data matrix; According to the height data matrix, the etching depth, surface roughness and etching residue evaluation indexes are calculated; Among them, the etching depth is determined by the height difference between the reference surface and the etching surface, the surface roughness is characterized by the arithmetic average roughness Ra, and the etching residue is obtained by setting a threshold to judge the discrete degree of height data; Measure the etching rates of samples of different material systems; Taking the chemical solution formula as input and the etching rate as output, a mapping relationship between the chemical solution formula and the etching rate is established; When the liquid formula changes, the corresponding etching rate can be quickly predicted; Calculate uniformity index and selectivity index; The uniformity index is obtained by the ratio of the standard deviation of the etching depth to the average value, which reflects the uniformity of the etching depth; The selectivity index is obtained by the ratio of the etching rates of two materials, reflecting the difference in etching ability of the solution formula on different materials; Use grey correlation analysis to determine the key process parameters that affect etching performance; The uniformity index and selectivity index are taken as parent factors, the solution formula and process parameters are taken as sub-factors, and the grey correlation between each sub-factor and the parent factor is calculated. The greater the correlation, the greater the influence of the factor on the etching performance. The results of grey relational analysis are used to guide the learning of intelligent optimization system; Taking uniformity index and selectivity index as optimization targets, liquid formulation and key process parameters as optimization variables, genetic algorithm is used to search for the optimal solution; When the intelligent optimization system completes an iteration, the experimental results are introduced into the sample set to update the knowledge base of the intelligent optimization system; Repeat the above steps continuously to achieve autonomous learning and iterative updates of the intelligent optimization system; As the optimization iterations proceed, the intelligent optimization system can summarize patterns from a large amount of experimental data, continuously improve prediction accuracy and optimization efficiency, and ultimately obtain the optimal solution formula and process parameter combination that takes into account both uniformity and selectivity.

Images